[Coco] Faster CoCo for George... Was: Re: OT: Gramical nitpiking >My CC3>CC4

[George Ramsower]

----- Original Message ----- 
From: "Joel Ewy"
> George Ramsower wrote:
>> ...
>> I would be tickled pink to have a 20mhz CC3 that would use my existing
>> hardware. I can easily get another keyboard and monitor. However, I
>> cannot easily replace all the stuff that plugs into a coco and the
>> software I've developed for it.
>> My CNC stuff would work faster and better also!
>>
>> This is selfish of me, I know.
>>
>>  I want MORE in a coco.
>>
>> More speed!
>>
>> Just simply.... more SPEED!
>>
>
> George, If you haven't already noticed Gary Becker's latest FPGA CoCo 3
> announcement in this list, you might want to take a look.  It appears
> after a first glance through the documentation that there is no CoCo
> compatible hardware expansion bus implemented in this design, so hooking
> your custom CNC motor controllers up to this beast may not be possible
> without some further work.  But it does run at about 20MHz, and the
> design is open.  So even if you don't feel like learning how to modify
> the design to do what you want, hooking up real CoCo peripherals to the
> FPGA board is likely to be something that many other users would want to
> do, so it may well be something that others will work on for you.  It is
> at exactly this point that James Jones will chime in and remind us that
> many FPGAs don't play nice with 5V logic.  :)  So making a CoCo
> expansion bus for this design may turn out to be non-trivial.  Still,
> this is a major step in the right direction towards a faster and more
> capable CoCo for the future.
>

 Would this sort of thing require a "Wait State" for the slower peripherals 
such as floppy controllers, RS232 cards and the like? Even my CNC stuff 
would be too slow, I think. However, I bet there are faster 74xxx chips out 
there that could replace the ones I'm using in my I/O board that connects to 
just about everything I do on that coco. It's just an address decoder with 
four input buffer chips and four output latches. Lately I've been thinking 
of adding more because, I may need more as this project continues.
 As with any computer project, the need for speed, memory and more speed 
increases with progress.
 I'm thinking a binary up/down counter on the stepper motor drivers would 
give me a REAL indication of their positions and any errors in math that may 
put them in the wrong place can be corrected on the fly, giving more 
reliable accuracy over the long haul.
 Of course, this will require more coding, more CPU time and thusly 
requiring more speed. Ugh.

George